octave-dev

[OctDev] octave-forge: inline problem

[John W. E. <jw...@be...>]

On 11-Jan-2003, Peter Brinkmann <bri...@ma...> wrote:

| Hi!
| I just tried Octave 2.1.42 with octave-forge under Windows 2000, and I've
| got a problem with the function inline.m from octave-forge. Here's
| what happened:
| 
| >> f=inline('x.*x','x')
| parse error near line 32 of file /usr/share/octave/2.1.42/m/miscellaneous/inline.m
| 
| >>>     argstr = ["x,",sprintf ("P%i,",1:n)];
|                                        ^
| error: `inline' undefined near line 1 column 3
| 
| I'm using the Octave binaries from Andy Adler's website.
| Any help would be appreciated.

This seems to work fine with my current sources and the current
version of inline.m from the octave-forge CVS archive, at least when I
start Octave without any arguments.  But it looks like you might be
using --traditional?  Then it also fails for me.  Ah, the bug is the
difference in parsing spaces inside [ ].  One fix is:

   argstr = ["x,",(sprintf ("P%i,",1:n))];

Another is to remove the space between sprintf and ( in that
statement.

jwe

Re: [OctDev] octave-forge: inline problem

[Andy A. <ad...@nc...>]

John W. Eaton wrote:

> But it looks like you might be
>using --traditional?  Then it also fails for me.  Ah, the bug is the
>difference in parsing spaces inside [ ].  One fix is:
>
>   argstr = ["x,",(sprintf ("P%i,",1:n))];
>
>Another is to remove the space between sprintf and ( in that
>statement.
>  
>
Yes. The problem is from --traditional.
I've just checked in this change.

andy

Re: [OctDev] octave-forge: inline problem

[John W. E. <jw...@be...>]

On 11-Jan-2003, Andy Adler <ad...@nc...> wrote:

| John W. Eaton wrote:
| 
| > But it looks like you might be
| >using --traditional?  Then it also fails for me.  Ah, the bug is the
| >difference in parsing spaces inside [ ].  One fix is:
| >
| >   argstr = ["x,",(sprintf ("P%i,",1:n))];
| >
| >Another is to remove the space between sprintf and ( in that
| >statement.
| >  
| >
| Yes. The problem is from --traditional.
| I've just checked in this change.

There is also another problem that Peter pointed out to me another
message, but it is really a bug with Octave.  If implict_str_to_num_ok
is nonzero, then isnumeric ('some string') will return true, because
it is possible to use 'some string' as a numeric value (with
conversion) at least in some contexts.  But I think it should not
really return true.  I'll fix that for 2.1.44.

jwe

Re: [OctDev] octave-forge: inline problem

[John W. E. <jw...@be...>]

On 11-Jan-2003, John W. Eaton <jw...@be...> wrote:

| There is also another problem that Peter pointed out to me another
| message, but it is really a bug with Octave.  If implict_str_to_num_ok
| is nonzero, then isnumeric ('some string') will return true, because
| it is possible to use 'some string' as a numeric value (with
| conversion) at least in some contexts.  But I think it should not
| really return true.  I'll fix that for 2.1.44.

Sorry, this fix didn't make it in 2.1.44, but it should be in 2.1.45.

jwe